CEOInterviews.AI
Start App
Gene Munster
Gene Munster
Managing Partner, Deepwater Asset Management
Watch

Dan Ives: The AI Crash Fear Is Completely Overblown

📅 Oct 25, 2026 Finance Deep Dive 14 MIN 39 VIEWS 7 SEGMENTS · 2 SPEAKERS
Is the AI stock sell-off a warning sign—or a buying opportunity? Gene Munster explains why he believes the AI revolution is still in its early stages, pointing to Meta’s AI assistants, Tesla’s autonomous driving technology, Nvidia’s physical AI ambitions, and expanding AI infrastructure. But with massive capital spending and questions about monetization, can the biggest technology companies turn AI investment into lasting revenue growth? Watch until the end for the key risk investors should not ignore. Chapters 00:00 — Is the AI sell-off just a pause? 00:45 — Meta’s AI ambitions 01:30 — Tes...

What Gene Munster said

Written from the verified transcript and checked against it. Every figure links to the moment it was said.

Gene Munster said the AI sell-off is overblown because the market is ignoring how early AI still is. He cited Meta's 10 million daily active users of Muse, Tesla's FSD attach rate rising from 15% three years ago to 70% now, Jensen's talk on physical AI, and SpaceX's plan to add 10 gigawatts of compute next year. Munster argued that personalized bots and agents use 10 to 15 times more tokens than typical chatbots, and physical AI is a second major driver. He dismissed concerns about overbuilding, saying companies spend out of belief, not just competition. Munster predicted next year's street estimate of 38% capex growth is too low, with the real number likely 70% or greater in calendar '28, and street's 11% growth is likely 25% to 40%. He called stocks 'great buys' on this nervousness.

Key takeaways

  1. Munster called the AI sell-off 'academic nervousness' and said stocks are 'great buys' at current levels.
  2. He predicted real capex growth will be 70% or greater in calendar '28, versus street's 11%.
  3. He forecast real capex growth of 25% to 40% next year, versus street's 38%.
  4. He said personalized bots and agents take 10 to 15x the token usage of typical chatbots.

Numbers and commitments

FigureWhat it refers toTypeAt
10 million daily active users of Meta's Muse metric 3:07
1% autonomy's share of total ride-sharing miles today metric 3:07
70% current FSD attach rate metric 3:07
10 gigawatts SpaceX compute to be added next year metric 3:07
3 SpaceX compute at end of this year in gigawatts metric 3:07
10 to 15x token usage of personalized bots versus typical chatbot metric 4:01
38% street's expected capex growth next year guidance 5:54
70% or greater real capex growth in calendar '28 guidance 5:54
11% street's expected growth for calendar '28 guidance 5:54
25 to 40% real growth for calendar '28 guidance 5:54

Chapters

  1. 0:00Reading the AI sell-off
  2. 3:07Early AI evidence from Meta, Tesla, SpaceX
  3. 4:01Why the market is wrong
  4. 5:08Responding to the transition period concern
  5. 5:54Capex growth predictions
  6. 7:13Defending spending against overbuild fears

Questions asked in this interview

3
  1. 2:33The market is getting a little bit twitchy about things that seem late cycle, at least on the capital markets side of AI. How are you reading it?
  2. 5:08So the question here is, do you buy the stock on the hope that this transition period is shorter than what we might think, that that gap will be bridged to what they spend in the monetization part of the phase?
  3. 7:13I mean, Gene, I'm not sure you necessarily have to conclude that Mark Zuckerberg and his CEOs are incompetent, but you can observe that all of the incentives of capitalism in these environments kind of drive you toward overbuild, right?
Mike 2:33 ↗
Mike. All right. Thank you. Yeah. The stock was down actually 6.4% in the regular session, so just gaining back a fraction of that so far. So could this tech sell-off be just a pause or the start of a deeper reversal? Joining us now is Deepwater Asset Management managing partner Gene Munster. Gene, good to see you. I know you're kind of fond of saying that we're very early with AI as it evolves. The market is getting a little bit twitchy about things that seem late cycle, at least on the capital markets side of AI. How are you reading it?
Gene Munster 3:07 ↗
Well, I'm reading what we heard this week, and I think it's important context to the big sell-off here. We heard from Meta, where they have 10 million or so daily active users of Muse, and Zuckerberg is basing the whole foundation of the company around that next year. That says something that is small that he believes will be very big — multibillion of users on a daily basis in the near term. We, of course, heard from Tesla and FSD. Today, autonomy represents about 1% of total ride-sharing miles. The attach rate for FSD has gone from about 15% three years ago to 70% today. Something bigger is going on, still nascent. We heard from Jensen today. We heard from Jensen at GTC in Berlin, and they talked about the power of physical AI. And we heard, of course, from SpaceX and this prediction by Elon that they're going to add 10 gigawatts of compute for next year. They'll be at about three at the end of this year.
So what do I make of all this is that if you believe that these leaders are competent, you should, I think, imply that we are still early. And specifically, what's happened — the difference that's happened in the last week versus the last three years — is these two phenomena around personalized bots and the agents, and them taking 10 to 15x the token usage versus a typical chatbot. And then separately, what can happen around physical AI. Those two collectively are much bigger than what we've seen over the past few years. And so there's a massive disconnect between what we're hearing loud and clear from these companies about how transformative this is going to be and how early this is going to be, and the market's academic unwillingness to accept that. And I think at the end of the day, truth will prevail. And I think the truth is this: that the world will be reoriented around AI in ways that we fully can't comprehend, that are beyond what I think the high expectations are today. And these stocks are ultimately great buys down here on this nervousness, this academic nervousness.
Mike 5:08 ↗
I understand the case, Gene. You make a very compelling one. But there is that period in between here where we get from, you know, 10 million users, which is a drop in the bucket compared to all the users across the Meta platforms, right? So, I mean, there is potential there, but it is a very small number, and getting up to the number of a billion, which is typically where Mark Zuckerberg says that's when you start monetizing things. There's a long runway. And in the meantime, the compute costs are very high. So the question here is, do you buy the stock on the hope that this transition period is shorter than what we might think, that that gap will be bridged to what they spend in the monetization part of the phase?

3 more exchanges in this transcript

Sign in free to read the rest of this interview. No card required.

Sign in to read the full transcript

Cite this transcript

APA, MLA, BibTeX
APA

Munster, G. (2026, October 25). Dan Ives: The AI Crash Fear Is Completely Overblown [Interview transcript]. Finance Deep Dive. CEOInterviews.AI. https://ceointerviews.ai/interview/3131103/

MLA

Gene Munster. "Dan Ives: The AI Crash Fear Is Completely Overblown." Finance Deep Dive, 25 Oct. 2026. Transcript, CEOInterviews.AI, https://ceointerviews.ai/interview/3131103/.

BibTeX 
@misc{munster2026_3131103,
  author       = {Gene Munster},
  title        = {Dan Ives: The AI Crash Fear Is Completely Overblown},
  howpublished = {Interview transcript, Finance Deep Dive. CEOInterviews.AI},
  year         = {2026},
  month        = {oct},
  url          = {https://ceointerviews.ai/interview/3131103/},
  note         = {Speaker-attributed transcript with timestamps}
}